BRAIN INJURY,
2001,
VOL.
15,
NO.
9, 787± 800
Comparison of pocket-computer memory aids
for people with brain injury
P ATRICIA WRIGHT{, NICK ROGERS‡,
CHRISTINE HALL‡, BARBARA
WILSON}{, JONATHAN EVANS‡},
HAZEL EMSLIE‡ and CHRISTINE
BARTRAM‡
{ School of Psychology, Cardiff University, Cardiff, Wales, UK
‡ MRC Cognition and Brain Sciences Unit, Addenbrooke’s Hospital, Cambridge, UK
} Oliver Zangwill Centre, Ely, Cambridgeshire, UK
(Received 21 September 2000; accepted 24 January 2001 )
Two styles of pocket computer memory aid were compared as support for people who had sustained
non-progressive, closed-head brain injury. A purpose-designed interface provided a diary with auditory alarms, a notebook and links between diary entries and specific notepages. One computer had a
physical keyboard, the other did not. Twelve adult volunteers were loaned each computer for 2
months, with a 1 month gap between, in counterbalanced order. It was found that all participants
could use the memory aids, and most (83%) found them useful. Little customizing was needed, but
amount of use varied widely. Predictors of usage included use of other reminding systems before
joining the project, and speed in calculator addition which may reflect working memory. High users
preferred the computer with a physical keyboard; low users made more entries with the palm-size
computer. These data highlight the need to distinguish ability to use from willingness to use.
Introduction
Memory aids such as diaries and address books are essential to daily life for most
people. There are recent reviews of the benefits offered by many kinds of memory
aids [1], and overviews of the range of electronic aids that are potentially beneficial
in the context of rehabilitation after brain injury [2]. For people with cognitive
impairments, conventional forms of memory support may not be enough. Paperbased aids require planning skills, which may be impaired. They also make heavy
demands on memory. People must remember to enter items into the diary, then
remember to look in the diary for forewarning of future activities, and remember to
carry out these activities at the appropriate time. Electronic memory aids can reduce
some of this memory load, e.g. by sounding an alarm when an activity needs to be
done and inserting repeated entries, and so can offer assistance to people with
memory problems in maintaining their independence and enhancing their quality
of life.
Correspondence to: Patricia Wright, School of Psychology, Cardiff University, PO Box 901,
Cardiff CF10 3YG, Wales, UK. e-mail: [email protected]
Brain Injury ISSN 0269± 9052 print/ISSN 1362± 301X online # 2001 Taylor & Francis Ltd
http://www.tandf.co.uk/journals
DOI: 10.1080/02699050110045161
788
P. Wright et al.
For patients with memory loss arising from brain injury it has been established
that paging systems can be highly beneficial [3]. So, too, can pocket computers [4]
with the advantages of greater flexibility in changing appointments together with
more scope for tailoring the look and feel of the interface to specific user needs.
People who have sustained brain injury will differ in the severity of their memory
loss and the impairment of other cognitive functions such as planning and organizing ability. Consequently, it is desirable that the interface of a computer-based
memory aid, together with the range of functions made available, can be easily
customized to suit individuals. Ways in which this might be done include simplifying how appointments and other text are entered into the aid [5]. The risk of
confusion, for example between entering reminders in an appointment diary or
in an electronic notebook, might be reduced by removing some of the functionality
or inserting a dialogue sequence that helps the user enter information in the appropriate place. The main goal of the project reported below was to see whether an
interface could be designed for memory aids on pocket computers, including an
appointment diary and a notebook, that could be easily mastered by people with
memory problems.
Pocket computers differ in whether or not they provide a physical keyboard. A
keyboard may facilitate text entry, and so encourage greater use, or it may make the
machine seem off-putting and too technical for participants with little or no experience of using computers. The alternative is to display an on-screen keyboard where
letters can be tapped by a pen. This enables the computer to be smaller and lighter,
but the keyboard reduces the display area of the already small screen. Entering text
via these keyboards is known to be error-prone, but people may not be bothered by
this when they are writing for themselves [6]. As the technologies of information
and communication develop, the convergence of pocket computers and mobile
phones may solve the portability problem whilst retaining the availability of the
keyboard. It, therefore, is very timely to ascertain whether people with memory
problems would be helped or hindered by use of a physical keyboard. This project
compares people’s attitudes to and use of pocket computer memory aids when
essentially the same interface is made available on a handheld computer with a
keyboard and on a palm-size computer having only a touch-screen keyboard. It
is difficult to predict which will be better suited for people who have sustained a
brain injury.
Method
The comparison of the two computers was a within-subject design, with participants using each machine for 2 months, with a gap of 1 month between machines.
The order of machines was counterbalanced across participants. Everyone followed
a 22-week protocol, in which a memory diary was kept during week 1.
The protocol
On the first visit, the researcher explained the purpose of the project, discussed the
memory aids currently being used, demonstrated the functionality of both computers, administered tests of visual acuity [7] and left a memory diary for daily
completion. In addition, participants did a short, timed test that combined manual
dexterity in using a small pocket calculator and short-term memory in entering
Pocket-computer memory aids
789
simple arithmetic calculations that were printed on a card. In the second visit, a
week later, the researcher demonstrated the memory aids on one of the pocket
computers and it was loaned to the participant. A follow-up visit took place in week
3. During a telephone interview in week 6, people were asked about their satisfaction with the computer and they were asked to use it to verify information known
to be in the computer (e.g. the researcher would ask `What time did we say I should
come on 22nd June?’). The computer was reclaimed after week 10 and participants
then had 4 weeks without a machine during which they were asked to record
instances when they felt they would have used the computer and whether they
managed successfully without it.
When the second computer was delivered, a short selection of psychometric
tests was given. These included the National Adult Reading Test (NART) [8], the
Speed and Capacity of Language Processing Test (SCOLP) [9], the Rivermead
Behavioural Memory Test (RBMT) [10] and the six elements subtest from the
Behavioural Assessment of Dysexecutive Syndrome (BADS) [11]. These psychometric measures would enable questions to be addressed concerning the cognitive
pre-requisites for benefiting from the memory aids, or the possible cognitive indicators of being unlikely to benefit. Participants were loaned their second pocket
computer for 8 weeks, and the protocol of visits and telephone interviews was
repeated as for the first computer. Details of actual usage were known from a
time-stamped computer log file of entries made. So, it was possible to distinguish
participants’ ability to use the full range of functions offered by the memory aid
(performance testing), and their willingness to use specific functions (computer log).
The memory aids
The computers
Both computers had grey-scale monochrome touch screens. The computer with a
keyboard was a Hewlett Packard (HP) 360 LX with 8 MB RAM, and a screen
resolution of 640 £ 240 pixels, occupying 154 £ 62 cm in landscape orientation.
The computer without a physical keyboard was a Casio E10 with 8 MB RAM, and
a screen resolution of 240 £ 320 pixels, occupying 80 £ 60 cm in portrait orientation. In spite of these size and format differences between the screens, it was possible
to display essentially the same information in a very similar format on both computer screens (see figures 1…a† and …b†).
The memory aids
Two memory aids, specifically designed for this target audience, were offered on
both computers. One was an appointment diary, the other a notebook. These aids
could be linked so that from a diary entry (e.g. Coffee with Alex) participants could
move swiftly to the notepage containing details about Alex. This would be useful if
Alex had to be contacted to rearrange the appointment, or if the diary user had
forgotten who Alex was. These links were set from the diary page and could be
made at any time, not only when first making the diary entry. For any diary entry,
people could set an alarm to ring and tick a box to repeat that entry daily or weekly.
Figure 2 shows the screen used for entering the details of an appointment into
the diary. The constraint of the 16 character slot size for each appointment was
790
Figure 1.
P. Wright et al.
The opening screen showing the similarity in the interfaces for accessesing the memory aids on the HP
(left) and the Casio (right).
Figure 2. The screens on each pocket computer after tapping the button `New Diary Entry’. For the palm-size
Casio (right), the on-screen keyboard displayed automatically. The HP had a clamshell style with the screen
in the top and a keyboard in the bottom casing.
overcome by the provision of a button labelled `More Text’, which enabled four
further slots to be displayed for the same timed appointment. The text in any extra
slots was shown on the diary page. People could use indentation via the space bar to
create a formatted display if they wished.
Navigation
Apart from text entry on the HP, all interactions with the memory aids were by
tapping the screen with the pen provided. There were two ways of moving
between diary pages. Arrows on the screen allowed people to move forward or
back one page at a time for as many pages as they wished. The button `Other Days’
displayed a vertical list of dates for the current week in the format Mon 21 June
1999, and arrows let people move to past or future weeks. The design decision to
show weeks rather than months was taken on the basis of an analysis of the frequency with which entries had been made in previous NeuroPage studies [2]. If this
was found inappropriate by present participants, they would have several opportunities to request changes during the project.
The computer notebook offered free text entry, together with the option to
label any page. A vertical list of labels could be displayed, and tapping a label in this
list opened that page in the notebook. Labels could also be included in a `Short List’
accessed from a button on the start-up screen which enabled swifter access to
important pages (e.g. the doctor) or any pages used often. Notepages without labels
could be found by using the arrows to go through the notebook page by page.
Pocket-computer memory aids
791
Participants
Five people who started the project discontinued for diverse, idiosyncratic personal
reasons. Three people did so after their first week with the pocket computer, all had
been loaned a Casio. Two people withdrew after having a pocket computer for 8
weeks, one had an HP and the other a Casio. The data reported below came from a
further 12 participants, six men and six women, with an average age of 34 years
(range 22± 54 years). Most had suffered traumatic brain injury, and six participants
had been under 30-years-old at the time of injury. The mean duration since brain
injury was 6 years (range 2± 12 years). Four participants had previously used
NeuroPage.
Exclusions
Exclusion criteria specified visual or motor handicaps that would have made it too
difficult for the volunteer to read the screen or to tap designated areas on the display.
Screening tests were conducted during the first visit. These included an assessment
of visual acuity based on the Lighthouse Near Vision Acuity Test (2nd edn) [7]. As
well as the smallest print that could be accurately read, participants were asked to
indicate the size they would find most comfortable for a newspaper or magazine. A
set of five cards was given, each requiring addition on a small calculator with a green
LED display. Two cards required the addition of six single digits; two involved six
double-digit numbers; one involved four three-digit numbers. The cards were
always presented in the same order. Nobody invited to participate was excluded
on grounds of visual impairment or poor motor control. Participants did not need
prior experience with computers, but people who were unwilling to use the aid,
either because of their attitude to computers or perception of their own memory
abilities, were excluded. Previous or present use of any other electronic memory aid
did not exclude participants.
Training
During the visit in week 2, participants were encouraged to find and read information in the diary and notebook, and to change some of that information. As figures 1
and 2 illustrate, the interface had been designed and pilot tested so that on-screen
choices left no uncertainty about which button to tap. Participants were also asked
to make their own new entries in both the diary and the notebook. They were
helped to put in the information they personally wanted and were given a manual,
comprised mainly of annotated pictures of the various screens, which was left with
them for the next 8 weeks. In week 3, they were visited and asked to demonstrate
how to find, change and enter details in both the diary and notebook, as well as
setting alarms and using the link between diary and notebook. Any forgetting or
idiosyncratic procedures that had evolved during the previous week were remedied.
Participants had two telephone numbers for contacting members of the project
team if they had any queries or if the computer went wrong.
792
P. Wright et al.
Results
All participants could use all the functionality available when asked to do so. The
design solution relied on people’s problem-solving skills when faced with a set of
unambiguous choices on the screen, rather than expecting them to remember
procedures. Nevertheless, it was found desirable to colour the pen tips for a few
participants, to increase the contrast with the screen so that they could more easily
see where they were tapping. In addition, the notebook interface was modified for
one participant to prevent conceptual errors in daily use, namely entering appointments into the notebook instead of into the diaryÐ an error triggered by prior use
of a paper notebook as a memory aid. The modification consisted of preventing any
changes to the notebook, but the participant’s partner was shown how to unlock it
to add new information. This solution was found acceptable to both participant and
carer. So, these data confirm that the project’s main objective was achieved. An
interface to memory aids on pocket computers can be designed that can be easily
used by people who have sustained brain injury. It needed almost no customizing to
meet the needs of these participants.
To address the issue of which style of pocket computer was most helpful for
participants, the findings will be presented separately for the qualitative data (questionnaires and interviews), and the performance data (computer logs of actual use).
Then, these will be related to the psychometric profiles. All statistical tests reported
below are two-tailed unless otherwise specified, but probability values just outside
0.05 are reported because in applied work such as this it may be more serious to
wrongly reject an intervention that could be beneficial than to accept the possibility
of a benefit that may not exist.
Qualitative data
The marks out of 10 given for features relating to ease of use of the diary and
notebook are summarized in table 1, which shows that the diary was considered
more useful than the notebook (pooling across computers, the mean rating for the
diary was 7.6 compared with 4.8 for the notebook, with 10 people showing a
difference in this direction and only two people the reverse (Sign test, p < 0:05).
The observation that diary ratings were higher for ease of use than for usefulness,
indicates that if people did not use the memory aids very much, this was not because
they found them difficult to use.
Evaluations of different diary features varied. For example, the alarms were felt
to be more useful (8.0) than the repeats (6.6). A difference in this direction was
shown by nine of the 10 people who reported a difference (Sign test, p < 0:02).
These mean figures mask considerable differences between individuals. On most
assessments, the Casio received higher ratings than the HP. Pooling across all ratings,
nine participants gave higher ratings to the Casio, only two people gave the HP
higher ratings (Sign test, p < 0:07). Nevertheless, several people commented on the
contrasting advantages of each computer: the lightness and portability of the Casio,
together with the value of having the alarm ring again if it was not responded to first
time, versus the louder alarm of the HP, its larger screen and convenient keyboard.
Attitudes to the computer having rechargeable batteries (HP) or not (Casio) varied.
A Casio advantage was also reflected in people’s explicit preferences when they
were asked at the end of their participation in the project which computer they
Pocket-computer memory aids
793
Table 1. Attitude towards ease of use as reflected in points out of 10 given
immediately after using each computer for 8 weeks (higher ˆ better)
HP
Casio
Mean
Diary
Ease of text entry
Overall ease of use
Usefulness of alarms
Necessity of repeats
Personal usefulness of machine
Mean
8.3
8.2
8.2
6.4
5.7
7.4
8.5
8.5
7.8
6.5
7.3
7.7
8.4
8.4
8.0
6.6
6.5
7.6
Notebook
Ease of labelling notes
Ease of linking to notes
How often notes used
Usefulness of short list
Mean
5.6
4.3
4.7
2.7
4.3
7.1
5.3
5.2
3.5
5.3
6.4
4.8
4.9
3.1
4.8
would prefer to keep. Seven people wanted the Casio, only three would keep the
HP, and two people had mixed preferences. Although not statistically significant,
these preferences are consistent with the pattern of data from the attitude ratings
shown in table 1, suggesting that the Casio was preferred to the HP.
One indication of the usefulness of the pocket computer memory aids to some
participants was that eight people said they might purchase a commercial pocket
computer in the future. These people accepted the offer of a list of currently
available machines, although it was pointed out that the diary and notebook they
had been using had been specially developed and were not available on commercial
machines. Project team members provided further assistance to any participants who
went ahead and purchased commercial machines, and that experience underscored
the desirability of having a purpose-designed interface. Prior to joining the project,
some participants had given up trying to master a commercial organizer.
During the gap between machines, participants were asked for 1 week to keep a
note of events that they would have entered into the electronic diary and to say
whether or not these had been successfully remembered. Only seven people made
entries in this record (mean number of entries was 11, range 7± 24), and only three
people recorded forgetting things they would have entered into the diary. A harsh
interpretation of these data would be that only 25% of participants benefited from
the electronic memory aids. An alternative possibility is that these records are
unreliable. Certainly, there was evidence from the handwriting that some records
were completed at a single sitting (e.g. just before the delivery of the new computer)
rather than throughout the week. Moreover, it is plausible that some activities that
were forgotten did not get recorded because they were still not remembered at the
time of completing the record, but they might have been prospectively entered into
the diary.
Use of the memory aids
During the visit in week 3 and the telephone interviews in week 6, all participants
could demonstrate the ability to use the electronic diary and the computer
794
P. Wright et al.
notebook to find information, make new entries and edit existing entries. They
could also make links between the diary and the notebook, and they could label
pages in the notebook. Therefore, measures of the use of specific functions as
recorded in the computer logs reflect choices made by participants rather than
limitations on that choice imposed by memory ability.
Frequency of use
The computer unobtrusively provided a detailed record of entries made in the diary
and notebook. However, this understates usage because participants may also have
read entries on the diary page, e.g. to check times of activities later in the day, or in
the notebook several times during the day. Participants varied widely in the use they
made of the memory aids. To facilitate comparison between the two computers, the
diary data were pro-rated to a duration of exactly 8 weeks, even though some
people had the machine for a few days longer (e.g. because of holidays or the
need to change appointments). In addition, entries relating to recharging the HP
over-night were excluded from the analysis because they would inflate the figures
for this machine relative to the non-rechargeable Casio.
The diary enabled people to enter repeated events by tapping a `Daily’ or
`Weekly’ button. Use of the diary can, therefore, be expressed in two ways.
Total entries includes repeats and corresponds to the number of diary entries that
would be read by a participant. In contrast, new entries excludes repeats after their
first occurrence, and corresponds to the number of different entries written by
participants. To illustrate how these measures may diverge, compare a person
who made one entry but repeated it daily for a week (7 total entries, 1 new
entry) with a person who made four entries relating to different activities during
that week (4 total entries, 4 new entries). The first person read more, but the second
person wrote more. An indication of the independence of these two measures is
given by the non-significant correlations between numbers of total and new entries
for each machine (Casio r ˆ 0:24, HP r ˆ 0:37). Where appropriate, both figures
are given in the data summaries below. When these two measures converge on the
same interpretation, for example in comparing the two pocket computers, this
indicates there is no `distortion’ due to repetition effects.
The data summarized in table 2 for an 8 week period suggests that participants
made approximately three total entries per day, with one new entry per day. On
average, both more diary entries and more notebook entries were made with the
Casio than the HP, but there was wide variation in frequency of use with both
machines. This variation seemed attributable to the participants rather than the
Table 2.
Entries made in pocket computer diary and notebook during an 8 week period
Diary
Total entries
Mean
SD
Range
n using Casio more than HP
Notebook
New entries
Note pages
Casio
HP
Casio
HP
Casio
HP
155
73.7
78± 337
7
146
134
10± 429
55
36.2
16± 135
8
53
23.2
8± 139
12.3
11.9
1± 43
6
10.2
8.1
0± 27
Pocket-computer memory aids
795
pocket computers, since people who were high users of one machine tended to be
high users of the other. The correlation between the two pocket computers was
significant both for total entries …r ˆ 0:82, p < 0:01) and new entries …r ˆ 0:71,
p < 0:01). So, for some analyses it will be adequate to pool across the computers.
Relation between computer and frequency of use
In order to establish whether the relative use of each pocket computer was similar
for high and low frequency users, the data were pooled across machines and participants divided into two subgroups. People in the high use subgroup made more
than 300 total entries (n ˆ 5, mean 496), and tended to make more total entries
with the HP (282) than the Casio (211) (Student’s t ˆ 2:48, df ˆ 4, p < 0:07). In
contrast, low users made less than 200 total entries (n ˆ 7, mean 174), but made
twice as many entries with the Casio (119) as with the HP (56) (Student’s t ˆ 4:97,
df ˆ 6, p < 0:01). This suggests that low frequency users were put off using the
pocket computer when it had a physical keyboard, whereas those entering more
text found the keyboard useful.
This pattern of the difference between the computers varying with frequency of
usage could be seen in the analysis of new entries, although it was not statistically
significant. Participants were divided into those who made more than 100 new
entries when pooling across machines (n ˆ 6, mean 168) and those who made less
than 100 (n ˆ 6, mean 58). For people in the high frequency subgroup, there was
little difference in the use of each computer (HP 85, Casio 83), with three people
making more new entries with the HP and three making more with the Casio.
Amongst the low frequency users, there was a tendency shown by five of the six
people for the Casio to be used more than the HP (HP 21, Casio 27, Student’s
t ˆ 0:86, df ˆ 5, n.s.). These data are consistent with the suggestion that the physical keyboard of the HP may have been found off-putting by low frequency users.
Table 3 summarizes the use made of additional functions for both the electronic
diary and computer notebook. The only statistically reliable difference between the
pocket computers was the proportionally greater use of alarms for diary entries
when using the HP (Student’s t ˆ 2:38, df ˆ 11, p < 0:04). This suggests that,
although overall the HP had fewer diary entries, these entries were critically time
related. This may serve as a caution against too simplistic an interpretation of the
quantitative data. A memory aid that performs a safety-net role of alerting the user
to critical events can be found very valuable even if that role is not required often.
Table 3.
Mean percentage of diary and notebook entries having additional functions
Diary entries
Notebook entries
Repeat
Casio
HP
Mean
Alarm
% Tot
Links
% New
% Tot
% New
More
text
% with
labels
% on
shortlist
68.9
82.6
74.7
7.5
6.7
7.1
60.9
50.9
55.9
14.1
13.5
13.8
16.5
5.6
11.5
94
80
87
49
58
54
796
P. Wright et al.
Psychometric profiles
The variability in the frequency of using the memory aids provides an opportunity
to consider whether any of the psychometric measures correlated with frequency of
diary usage. The memory diaries were not always fully completed during week 1,
but eight people completed the diary for at least 5 days. The mean number of
memory problems reported per person was 52.8 (range 3± 152). Ten participants
reported memory problems in five or more different contexts (range 2± 20).
However, there was no relation between total number of memory problems
reported and the number of entries made in the pocket computers …r ˆ 0:32†. It
was noticed that people who withdrew from the study had reported fewer memory
problems during the first week than the average participant. (For people who
withdrew, the mean number of all problems ˆ 12.8; mean number of different
problems ˆ 5.5.) If this indicated satisfaction with their current memory performance it would explain why they felt less need for memory aids.
The tests of visual acuity and simple arithmetic using a calculator had been
intended as a means of filtering volunteers who might have problems using the
pocket computers. However, correlations with the number of diary entries made on
both computers showed that time taken to complete the additions correlated with
total entries …r ˆ 0:69; p < 0:02† but not the number of new entries …r ˆ 0:01†.
This indicates that people who were slower at using the calculator, perhaps because
of working memory problems rather than manual dexterity problems, tended to
have more total entries in their diaries. There was no similar relation between the
use of computer diaries and visual comfort levelÐ this measure was chosen because
it yielded a wider range than visual acuity (total entries r ˆ 0:24, new entries
r ˆ 0:01).
Sometimes, participants had already done one or more of the tests from the battery
used here, and so their scores were now a little higher than they had been on previous
occasions. That is to say the measures reported here may be over-estimates of ability in
a few instances. A summary of the distribution of participants across the various tests is
given in table 4. There were no significant correlations between any single psychometric measure and either the number of total entries or new entries made.
Table 4.
Average or above
Slightly impaired
Severely impaired
Summary of psychometric data on participants
NARTa
FSIQ
RBMTb
STDP
SCOLP
SWc
SCOLP
SCd
BADSe
8
3
1
2
6
4
7
1
4
5
3
4
7
3
2
0.06
0.33
70.45
0.26
Correlation with combined diary entries on both machines
(Need r > 0.57 for significance at p < 0.05 two-tailed)
Total diary entries
0.50
0.52
0.19
New diary entries
0.05
0.28
0.44
a
National Adult Reading Test [8] indicates intelligence prior to the injury.
Rivermead Behavioural Memory Test [10] indicates level of memory impairment.
c
Spot the Word test from the Speed & Capacity of Language Processing Test [9].
d
Speed of Comprehension test from the Speed & Capacity of Language Processing Test [9].
e
The Six Elements subtest from Behavioural Assessment of the Dysexecutive Syndrome [11].
b
Pocket-computer memory aids
Table 5.
797
Relation of estimate of pre-morbid IQ to the number of diary entries made
Combined entries
Subgroup
Hi …n ˆ 5†
Med …n ˆ 4†
Lo …n ˆ 3†
Table 6.
Total diary entries
Mean FSIQ
Total
New
HP
Casio
115
103
91
412
288
130
101
198
33
219
139
32
193
149
98
Relation between prior use of memory aids and use of the pocket computer diaries
Total entries
New entries
Combined
Prior use of memory aids
HP
Cas
HP
Cas
HP
Cas
Often …n ˆ 6†
Seldom …n ˆ 6†
189
120
214
77
61
49
82
24
404
197
143
73
Because of the border-line correlation with FSIQ, which estimates pre-morbid
intelligence, the use of each machine was explored further by grouping participants
into one of three bands (Hi ˆ 121± 111, Med ˆ 110± 97, Lo ˆ 96± 87). Table 5
shows that the total number of entries decreased across these three FSIQ bands,
and the biggest difference between the two computers was for those people with
the lowest FSIQ scores where the HP was used very little. Caution is needed in
interpreting these data since there were only three people in this FSIQ band, but if
this is a reliable effect it should replicate in the ongoing study using a similar
interface to access a greater range of memory aids on pocket computers.
Another possible critical factor influencing the amount of use made of the
pocket computers could be participants’ prior use of reminding systems. On the
basis of interview data, participants were classified into two groups; those who were
systematic users of memory aids (e.g. diary, calendar, filofax or notebook) prior to
joining the project, and those who were not. People were encouraged not to
abandon their other aids, since the pocket computers would only be available for
the project duration. It might have been expected that people accustomed to using
other aids might make less use of the pocket computers, but table 6 shows that this
was not the case. People accustomed to using memory aids, of whatever type,
tended to make more use of the pocket computers, although the differences
between these subgroups of prior organizers were not statistically reliable on a
two-tailed test (total entries Student’s t ˆ 2:04, df ˆ 10, p < 0:07; new entries
Student’s t ˆ 1:73, df ˆ 10, p < 0:06, one-tailed). Sometimes participants used
the old and the new aids in conjunction with each otherÐ e.g. setting an alarmed,
repeated reminder on the computer to check in their paper diary.
Hardware problems
Two HPs failed and were replaced by the manufacturer. One Casio screen developed a small permanent black patch which might have come from a knock. The
diary interface was designed to prevent easy access to the software already installed
798
P. Wright et al.
by the manufacturer on the computers. Nevertheless, games were found and played
by some participants, and these occasionally caused problems, requiring resetting the
machine, which was less straightforward on the HP. The Casio batteries were not
rechargeable and for some participants, especially those who often used the backlight to enhance legibility, had to be replaced frequently (e.g. weekly). This problem was exacerbated by the manufacturer’s warning message about low battery
status which tended to come on earlier than necessary. There were occasional
problems with the Casio calibration (the relation between where the user tapped
and where the screen thought it had been tapped). However, there was no obvious
relation between a participant’s experience of hardware problems and their willingness to use the aid.
Discussion
The main project objective of designing a memory aid that could be used by people
after brain injury was achieved. Participants could enter and retrieve information
from the memory aids when asked to do so and very little customizing of the
interface was needed to accommodate individual needs. With a brighter, colour
screen even the colouring of the pen tips may not have been needed. This success
was achieved by creating an interface to which people could apply their problemsolving skills to select amongst the choices displayed, rather than having to remember specific details of how to make entries in the diary or set alarms. This is not to
suggest that the interface was perfect for everyone. The debriefing interviews elicited a variety of requests, usually for additional functionality, some of which are
being followed up in this ongoing project.
In case it was needed, the research team had created an even more supportive
interface, where only one element of making a diary entry was possible on each
screen, and the actions needed for doing this were explicitly prompted on the
screen. But, no participants needed this additional guidance. The attitude ratings
confirmed that people found it very easy to use the diary on both computers.
However, the variation in the frequency with which the aids were used suggested
that there were a number of other critical factors determining how much the diary
would be used.
One potentially important factor could be the patients’ insight into their need
for memory assistance. People may not bother to enter reminders for events they
think they will remember. Another possibility is that a more general lack of motivation underlies the reluctance to use the aid. This possibility is being addressed
within this ongoing project by introducing other activities for which the pocket
computer can be used (e.g. games). If people are encouraged to use the computers
more often, they might find themselves more willing to enter reminders in the diary
while they have the pocket computer in their hands and switched on. A third
possibility, and one supported by the data in table 6, is that use of the memory
aids relates to time management and scheduling skills. People who tend to plan their
day ahead in some detail may be those who use the memory aids often. This
suggestion is consistent with reports that people who used a variety of remembering
strategies prior to their injury were likely to use more after the injury than people
who premorbidly used few external reminding strategies [11]. Others have
reported finding a similar relation [13], and this may point to the value of training
Pocket-computer memory aids
799
people in how to use diaries and memory aids during rehabilitation from brain
injury.
This relation between use of the pocket computers and prior use of memory aids
might reflect the role of familiarity-plus-perseveration , with people preferring to
continue managing their daily routines after the accident by whatever methods they
had been using before. However, there was some evidence for a knowledge or skills
deficit contributing to low use of the aids. Amongst the participants in the project,
not everyone seemed to understand how to use reminding devices to their best
advantage. For example, one participant had a fixed shopping list that was looked at
only in the shop, where decisions were made about whether specific items were
needed this week. If remembering things is a problem, then this strategy may not be
as safe as checking the items before leaving home. It is known that interface factors
may miscue people about the cognitive demands of available options [15], but that
seems unlikely to have been a factor in the present study. An appropriately designed
interface can ensure procedural adequacy, but it is not so easy to help participants set
relevant goals for using memory aids.
The present study has provided data on the relative suitability of two pocket
computers, one with and one without a keyboard. These data strongly suggest that
different pocket computers suit different participants. People who used the memory
aids often, made greater use of the pocket computer with a keyboard (the HP). In
contrast, people who made few entries in the diary made greater use of the smaller
palm-size computer without a keyboard (the Casio). While it might be expected
that people who make more entries will value more highly the physical keyboard as
a convenient means of entering text, the strong antipathy for the HP among those
who made relatively little use of the diary and other memory aids is less easily
accounted for. The compromise of a palm-size machine with a detachable keyboard
might satisfy both high and low frequency users.
The benefits of a simplified interface for a pocket computer memory aid are
likely to extend beyond brain injured patients, for example to older people, many of
whom experience age-related memory loss [14]. Indeed, there is evidence that older
people have greater difficulty with remembering timed rather than untimed events
[16], and so should benefit from the auditory alarm coupled with a visual reminder
of the activity that the pocket computer provides.
Conclusions
This study has established that people with memory impairments resulting from
brain injury can use purpose-designed computer-based memory aids, comprising an
appointment diary, notebook and links between them. Several participants found
these aids of great benefit. The design solution is to present people with an unambiguous set of choices and then rely on their problem-solving skills rather than
expecting them to remember procedures. The amount of use made of the pocket
computers varied widely between individuals, and appeared related to people’s use
of other memory aids. Frequent users preferred the computer with a keyboard,
whereas infrequent users preferred the smaller, palm-size device. A compromise
such as a palm-size computer with a detachable keyboard might meet the needs
of both groups.
800
Pocket-computer memory aids
Acknowledgements
This study was made possible by grant PCD2/A1/215 from the NHS National
Programme for People with Physical and Complex Disabilities. The authors also
thank Professor Marslen-Wilson for accommodating the project within the MRC
Cognition and Brain Sciences Unit, and Hilary Williams of Cardiff University for
assistance with data analysis.
References
1. Herrmann, D., Brubaker, D., Yoder, C. et al.: Devices that remind. In: F. T. Durso (editor)
Handbook of Applied Cognition (Chichester: Wiley), pp. 377± 407, 1999.
2. Kapur, N.: Memory aids in the rehabilitation of memory disordered patients. In: A. D. Baddeley,
B. A. Wilson and F. N. Watts (editors) Handbook of Memory Disorders (Chichester: Wiley), 1995.
3. Wilson, B. A., Evans, J. J., Emslie, H. et al.: Evaluation of NeuroPage: a new memory aid.
Journal of Neurology, Neurosurgery, and Psychiatry, 63: 113± 115, 1996.
4. Kim, H. J., Burke, D. T., Dowds, Jr. M. M. et al.: Electronic memory aids for outpatient brain
injury: follow-up findings. Brain Injury, 41: 187± 196, 2000.
5. Masui, T.: An efficient text input method for pen-based computers. Proceedings CHI’98 (New
York: ACM Press), pp. 328± 335, 1998.
6. Wright, P., Bartram, C., Rogers, N. et al.: Text entry on handheld computers by older users.
Ergonomics, 43: 702± 716, 2000.
7. Lighthouse Near Vision Acuity Test (2nd edn). Lighthouse Low Vision Products, Long Island and
City, NY, USA.
8. Nelson, H. E.: National Adult Reading Test, 2nd edn (Berkshire: The National Foundation for
Educational Research), 1991.
9. Baddeley, A. D., Emslie, H. and Nimmo-Smith, I.: Speed & Capacity of Language Processing Test
(Bury St Edmunds: Thames Valley Test Co), 1992.
10. Wilson, B. A., Cockburn, J. and Baddeley, A. D.: Rivermead Behavioural Memory Test (Bury St
Edmunds: Thames Valley Test Co), 1985.
11. Wilson, B. A., A lderman, N., Burgess, P. W. et al.: Behavioural Assessment of the Dysexecutive
Syndrome (Bury St Edmunds: Thames Valley Test Company), 1996.
12. Wilson, B. A. and Watson, P.: A practical framework for understanding compensatory behaviour in people with organic memory impairment. Memory, 4: 465± 486, 1996.
13. Evans, J. J., Wilson, B. A., Needham, P. et al.: Who makes good use of memory aids? Paper
presented at the annual conference of the Society for Cognitive Rehabilitation, Birmingham, UK, 1999.
14. Wright, P., Lickorish, A. and Milroy, R.: Route choices, anticipated forgetting and interface
design for online reference documents. Journal of Experimental Psychology: Applied, 6: 158± 167,
2000.
15. Smith, A. D. and Earles, J. L. K.: Memory changes in normal aging. In: F. Blanchard-Fields and
T. M. Hess (editors) Perspectives on Cognitive change in Adulthood and Aging (New York: McGraw
Hill), pp. 192± 220, 1996.
16. Einstein, G. O. and McDaniel, M. A.: Retrieval processes in prospective memory: theoretical
approaches and some new empirical findings. In M. Brandimonte, G. O. Einstein, and
M. A. McDaniel (editors) Prospective Memory: Theory and applications (Hillsdale, NJ: Lawrence
Earlbaum), pp. 115± 141, 1996.